The Agent Skills Directory

[SAFE]: No malicious prompt injection or behavior override patterns were found in the instructions or reference files.- [SAFE]: The skill does not access sensitive local file paths (such as .ssh or .aws) or perform unauthorized network exfiltration; all examples use safe placeholders or local addresses.- [SAFE]: No obfuscation techniques, such as Base64-encoded commands, zero-width characters, or homoglyphs, were detected.- [SAFE]: All external libraries mentioned (e.g., Celery, Tenacity, Structlog, OpenTelemetry) are well-known, trusted industry standards for Python development.- [SAFE]: Database patterns shown in the background jobs reference correctly use parameterized queries, which is a key security measure against SQL injection.- [SAFE]: No privilege escalation commands (sudo) or persistence mechanisms (cron, shell profiles) were identified.- [SAFE]: The skill does not use the dynamic context injection syntax (!command) in its markdown files.

python-infrastructure